Automatic English Pronunciation Evaluation Algorithm Based on Sequence Matching and Feature Fusion

This article focuses on the question answering type of automatic scoring system for large-scale spoken English examinations and scores using a method called multifeature fusion. Three types features are extracted to score speech recognition text as research object. The three similarity, syntactic, phonetic. There nine distinct characteristics that describe relationship between examinee responses expert ratings. Manhattan distance is improved measure similarity in feature. Simultaneously, feature keyword coverage based editing proposed, phenomenon word variation fully considered, order provide examinees with an objective fair score. To obtain machine score, all were fused multiple linear regression model. experimental results demonstrate extremely effective scoring, performance unit equaling 98.4 percent performance.